Visual verification of historical Chinese calligraphy works

The problem of historical Chinese calligraphy verification is previously investigated by experienced artists, whereas this paper proposes some objective measures to bear the problem with evidences by analyzing the subtle discrepancies between the images of the suspicious and the genuine. First, features that characterize an individual calligrapher's writing style are extracted and modeled. When a suspicious comes, it is compared with the genuine in the reference database to detect problematic characters and to calculate total accepting probability. The efficiency of the algorithm is demonstrated by a preliminary experiment with 13274 images of calligraphy character.